Installation¶

Requirements¶

Supported Python versions are 3.10 through 3.14. NumPy, SciPy, and matplotlib are installed as base dependencies.

Installation via pip¶

Virtual Python environment¶

Create and activate a virtual environment before installing:

python -m venv .venv
source .venv/bin/activate

The activation command above is for POSIX shells. On Windows, use .venv\Scripts\activate for Command Prompt or .venv\Scripts\Activate.ps1 for PowerShell. Run deactivate to leave the environment.

Standard users¶

As a standard user, you can install the latest release via pip by

pip install iwopy

To install the unreleased development branch instead:

pip install 'iwopy @ git+https://github.com/FraunhoferIWES/iwopy@dev'

SciPy and the iwopy optimizers are included. Install an optional backend using its extra: cma, nevergrad, optuna, deap, pymoo, or pygmo. For example:

pip install 'iwopy[optuna]'

Use opt for all supported backends:

pip install 'iwopy[opt]'

The optimizer catalog contains a runnable setup for each optimizer. Missing optional packages produce an installation hint. The Nevergrad extra currently requires NumPy below 2.4; the catalog explains the upstream limitation.

Developers¶

The first step as a developer is to clone the iwopy repository by

git clone https://github.com/FraunhoferIWES/iwopy.git

Enter the root directory by

cd iwopy

Synchronize the development and test dependencies with

uv sync --extra test --extra dev

Install the pre-commit hooks with

uv run pre-commit install

Installation via conda¶

Conda installation¶

The iwopy package is available on the channel conda-forge. You can install the latest version by

conda install -c conda-forge iwopy

Development workflow¶

Use the repository’s uv workflow above for development, even when conda is used for other environments. See Testing for the quality gates.